Back to Home

A bit about DWDM 400 / 500G, 1Tb, about super channels and the future

dwdm

A bit about DWDM 400 / 500G, 1Tb, about super channels and the future

    Having stood next to the tested DWDM 100G, willingly or not, I began to read everything that came to hand and dug up some material on the prospects for the further development of DWDM technology. Gradually realizing the material, I will publish the results of awareness - maybe someone will be interested. Only part of the information is published here - the continuation will be later.
    In the near future, the developers promise us the possibility of transmitting information at speeds of 400 / 500G and even 1TB per wavelength.

    What is DWDM?



    I will outline in a nutshell. DWDM is a channel spectral multiplexing technology. What is the meaning of spectral multiplexing? Through one fiber optic cable, you can transmit more than one signal in parallel, using different wavelengths of light for this. There are several technologies of spectral compression — WDM, CWDM, DWDM, HDWDM. Technologies are distinguished by channel density, i.e. how far is one channel from another on the frequency grid.

    What systems are there now?



    Currently, DWDM and HDWDM systems are of most interest, since they provide the ability to transfer the largest amount of information through a single fiber optic cable. If we consider existing systems, we can distinguish technologies that provide data transfer at a speed of 2.5 / 10 / 40G on one channel. Total channels can be created up to 80 or up to 96, it all depends on the manufacturer of the equipment and what frequency grid it uses. Those. The total throughput of one optical fiber will be = (80 or 96) * (transmission rate per channel). For example, for a system with 80 channels and a transmission rate per channel of 10G, the maximum speed at which data can be transmitted through one cable = 10G * 80 = 800G
    Currently, most manufacturers are launching complexes on the DWDM 100G system that can transmit information at a speed of 100G / s per channel (lambda) and produce a maximum power of 9.6 Tbit / s. I stood a little while when testing these systems, which I already wrote about once on Habré (however, I had to hide the post, because testers were outraged - like I’m violating privacy) I was puzzled by the 100G problem and looked a bit further into the future. What is there?

    Prerequisites for the development of technology.



    I think it is no secret to anyone that technologies are developing in accordance with the requirements of the services market. If you look at the situation in general, you can see that in the future we are waiting for an avalanche-like increase in high-quality traffic transmission lines. First of all, as I see it, the 4th generation LTE networks that were actively built by mobile operators in our country will be “to blame” for this. If we consider that LTE, even in the truncated configuration that is supposed to be implemented in our country, assumes up to 80 Mbit downlink, making a discount on realities and dividing the menacing figure by 8, we get 10 Mbit for one mobile terminal. Which is on average 5-7 times more than the existing 3g with HSPA + offers. Yes, I know that completely different numbers are written in the standards, but in reality they look something like this. Unfortunately.
    So, if our mobile operators think, then they will understand that at the moment when each mobile phone starts to want the speed of 7-15 Mbit / s via the radio channel, the speed to transport networks will increase. And it will increase until LTE reaches its ceiling. The ceiling for LTE, as I recall, is designated at 400 Mbps on the downlink.
    At some point, the operators will have a question, how to further develop transport networks? Then you will either have to lay new optical lines, or install more modern equipment in order to better use the old ones. If we compare the cost of laying the lines and the cost (even of the newest DWDM complexes) - I have no doubt for a second what the operator will choose. Hence the relevance of the issue with the development of DWDM technologies to speeds exceeding 100G.
    Annual studies show that traffic is growing at about 40% per year.

    DWDM Systems with Speeds> 100G



    Well, here I come to the most interesting. How will systems with speeds> 100G per channel be organized, what will they be, etc. But all in order. As far as I managed to find out, the development stages of DWDM technologies are supposed to be up to 400, 500 Gbit per super-channel * and 1Tbit-a. No others were noticed.
    * Super-channel is defined as a logical combination of several physical channels, which on the client side is visible as a single channel with a bandwidth> 100G (400, 500 or 1000Gbps). Those. if at the input you have a client signal in 1Tb, you start it on a special Super-channel line card that processes the client stream in 1Tb and drives it through several carrier channels in a line. This is what the prototype 1Tb module from Infinera looks like:

    1Tb DWDM module from Infinera

    There are currently no approved standards defining how the super channel should work. Within the framework of standardization of the super-channel, the following work:
    1) Optical Internet Forum, is developing the possibility of expanding its developments in the field of 100G DWDM and including super-channel logic in them
    2) ITU-T in G.694.1 approved a frequency grid with a step option of 12.5 GHz
    3) ITU-T SG15, Question 11 - a research question has been opened on the development of new containers taking into account the capabilities of the 12.5 GHz frequency grid (developed by Infinera, Finisar, Verizon)
    Here ITU-T proposes to expand G.709 recommendation and an additional container in the frame structure, which will allow to correlate one OCh with n How many optical carriers. The diagram will now look like this:

    image

    4) IETF CCAMP develops the Generalized Label for Super-Channel Assignment on Flexible Grid 12.5 GHz, as well as the OSPFTE extension for supporting GMPLS in a 12.5 GHz grid (developed by Infinero and Verizon)

    Thus, in the case of the invention of transponder modules with the possibility of tuning through 12.5 GHz and with a transmission speed of 100G (against the existing 25GHz, and even then only with Infiner) we can get a 2 times gain in line capacity due to a denser frequency grid. If we assume that the channel will be 100G, then the approximate total speed of information transfer to the line can reach 32Tbit / s.

    The topic is new, there is not much information, most of the developments are not open for general use - so please treat me with leniency - it did not turn out completely and chaotically.

    Read Next